Ontology-Driven Pattern Selection and Matching in Software Design.

Design patterns are a meaningful technology for supporting the construction and modeling of software systems. Besides their use is related to the non-functional requirements fulfillment that is also an open challenge in the field of software design. In this work we propose a theoretical approach for modeling relationships and sequences of patterns and for modeling the taxonomy that relates patterns with ensured nonfunctional requirements for given application contexts. The approach is based on the use of Description Logics for modeling the domain of patterns and for reasoning tasks on the modeled domain. We developed a framework for supporting the architectural modeling phase and used it to verify the effectiveness of both the patterns conceptualization and the use of non-standard reasoning tasks for querying the pattern ontology.